I will use call IAudioFlinger::setMode API as a sample scenario to show how the Android IPC

system works. AudioFlinger is a service in the media_server program

Service Manager Run

service_manager provide service manager service to other process. It must be started before any

other service gets running.

int main(int argc, char **argv)

{

struct binder_state *bs;

void *svcmgr = BINDER_SERVICE_MANAGER;

bs = binder_open(128*1024);

if (binder_become_context_manager(bs)) {

LOGE("cannot become context manager (%s)\n", strerror(errno));

return -1;

}

svcmgr_handle = svcmgr;

binder_loop(bs, svcmgr_handler);

return 0;

}

It first open “/dev/binder” driver and then call BINDER_SET_CONTEXT_MGR ioctl to let

binder kernel driver know it acts as a manager. Then it enters into a loop to wait for any data from

other process.

void binder_loop(struct binder_state *bs, binder_handler func)

{

int res;

struct binder_write_read bwr;

unsigned readbuf[32];

bwr.write_size = 0;

bwr.write_consumed = 0;

bwr.write_buffer = 0;

readbuf[0] = BC_ENTER_LOOPER;

binder_write(bs, readbuf, sizeof(unsigned));

for (;;) {

bwr.read_size = sizeof(readbuf);

bwr.read_consumed = 0;

bwr.read_buffer = (unsigned) readbuf;

res = ioctl(bs->fd, BINDER_WRITE_READ, &bwr);

if (res < 0) {

LOGE("binder_loop: ioctl failed (%s)\n", strerror(errno));

break;

}

res = binder_parse(bs, 0, readbuf, bwr.read_consumed, func);

if (res == 0) {

LOGE("binder_loop: unexpected reply?!\n");

break;

}

if (res < 0) {

LOGE("binder_loop: io error %d %s\n", res, strerror(errno));

break;

}

}

}

Pay attention to BINDER_SERVICE_MANAGER.

/* the one magic object */

#define BINDER_SERVICE_MANAGER ((void*) 0)

BINDER_SERVICE_MANAGER is the registered handle for service_manager. The other process

must use this handle to talk with service_manager.

Get IServiceManager

To get an IServiceManager instance the only method is to call defaultServiceManager

implemented in IServiceManager.cpp.

sp<IServiceManager> defaultServiceManager()

{

if (gDefaultServiceManager != NULL) return gDefaultServiceManager;

{

AutoMutex _l(gDefaultServiceManagerLock);

if (gDefaultServiceManager == NULL) {

gDefaultServiceManager = interface_cast<IServiceManager>(

ProcessState::self()->getContextObject(NULL));

}

}

return gDefaultServiceManager;

}

gDefaultServiceManager is defined in libutil, so any program or library which included libutil will

have this symbol, it’s unique in one process. The first time gDefaultServiceManager is NULL, so

it will first get a ProcessState instance through ProcessState::self(). One process has only one

ProcessState instance. ProcessState will open “/dev/binder” driver for IPCThreadState use.

ProcessState::ProcessState()

: mDriverFD(open_driver())

Now we have an instance of ProcessState, let’s look at the getContextObject.

sp<IBinder> ProcessState::getContextObject(const sp<IBinder>& caller)

{

if (supportsProcesses()) {

return getStrongProxyForHandle(0);

} else {

return getContextObject(String16("default"), caller);

}

}

The board support binder driver, so we will get into getStrongProxyForHandle. (Handle 0 is

reserved for service manager, which will be explained later.)

sp<IBinder> ProcessState::getStrongProxyForHandle(int32_t handle)

{

sp<IBinder> result;

AutoMutex _l(mLock);

handle_entry* e = lookupHandleLocked(handle);

if (e != NULL) {

// We need to create a new BpBinder if there isn't currently one, OR we

// are unable to acquire a weak reference on this current one. See comment

// in getWeakProxyForHandle() for more info about this.

IBinder* b = e->binder;

if (b == NULL || !e->refs->attemptIncWeak(this)) {

b = new BpBinder(handle);

e->binder = b;

if (b) e->refs = b->getWeakRefs();

result = b;

} else {

// This little bit of nastyness is to allow us to add a primary

// reference to the remote proxy when this team doesn't have one

// but another team is sending the handle to us.

result.force_set(b);

e->refs->decWeak(this);

}

}

return result;

}

The first time b will be NULL, so the code will new an BpBinder instance. BpBinder is a base

proxy class for remote binder object.

BpBinder::BpBinder(int32_t handle)

: mHandle(handle)

, mAlive(1)

, mObitsSent(0)

, mObituaries(NULL)

{

LOGV("Creating BpBinder %p handle %d\n", this, mHandle);

extendObjectLifetime(OBJECT_LIFETIME_WEAK);

IPCThreadState::self()->incWeakHandle(handle);

}

IPCThreadState::incWeakHandle will add a BC_INCREFS command in output buffer.

void IPCThreadState::incWeakHandle(int32_t handle)

{

LOG_REMOTEREFS("IPCThreadState::incWeakHandle(%d)\n", handle);

mOut.writeInt32(BC_INCREFS);

mOut.writeInt32(handle);

}

Now getContextObject returns a BpBinder instance, it will be interface_cast to IServiceManager.

interface_cast is defined in IInterface.h. It will be extended as:

inline sp<IServiceManager> interface_cast(const sp<IBinder>& obj)

{

return IServiceManager::asInterface(obj);

}

Now let’s take a look at definition of IServiceManager.

class IServiceManager : public IInterface

{

public:

DECLARE_META_INTERFACE(ServiceManager);

/**

* Retrieve an existing service, blocking for a few seconds

* if it doesn't yet exist.

*/

virtual sp<IBinder> getService( const String16& name) const = 0;

/**

* Retrieve an existing service, non-blocking.

*/

virtual sp<IBinder> checkService( const String16& name) const = 0;

/**

* Register a service.

*/

virtual status_t addService( const String16& name,

const sp<IBinder>& service) = 0;

/**

* Return list of all existing services.

*/

virtual Vector<String16> listServices() = 0;

enum {

GET_SERVICE_TRANSACTION = IBinder::FIRST_CALL_TRANSACTION,

CHECK_SERVICE_TRANSACTION,

ADD_SERVICE_TRANSACTION,

LIST_SERVICES_TRANSACTION,

};

};

DECLARE_META_INTERFACE macro is defined as follows in IInterface.h. And it will be

extended to the following code:

static const String16 descriptor;

static sp<IServiceManager> asInterface(const sp<IBinder>& obj);

virtual String16 getInterfaceDescriptor() const;

As you can see, DECLARE_META_INTERFACE macro declares two functions which are

implemented by IMPLEMENT_META_INTERFACE macro in IServiceManager.cpp.

IMPLEMENT_META_INTERFACE(ServiceManager, "android.os.IServiceManager");

The code will be extended like this.

const String16 IServiceManager::descriptor(NAME);

String16 IServiceManager::getInterfaceDescriptor() const {

return IServiceManager::descriptor;

}

sp<IServiceManager> IServiceManager::asInterface(const sp<IBinder>& obj)

{

sp<IServiceManager> intr;

if (obj != NULL) {

intr = static_cast<IServiceManager*>(

obj->queryLocalInterface(

IServiceManager::descriptor).get());

if (intr == NULL) {

intr = new BpServiceManager(obj);

}

}

return intr;

}

So IServiceManager::asInterface will finally new a BpServiceManager instance and return it to

user. BpServiceManager works as a proxy for remote BnServiceManager. Any operation on

IServiceManager now actually is to call the corresponding virtual functions in BpServiceManager.

Summary:

This section gives out details in how to get a proxy object for remote object.

Assume you want to implement your own service IFunnyTest, you must do the following:

 Put DECLARE_META_INTERFACE(FunnyTest) macro in your interface header file.

 Put IMPLEMENT_META_INTERFACE(Funnytest, “your unique name”) macro in your

interface source file.

 Implement your own BpFunnyTest class.

Generate AudioFlinger Service

media_server program will start the AudioFlinger service. Here is the code.

int main(int argc, char** argv)

{

sp<ProcessState> proc(ProcessState::self());

sp<IServiceManager> sm = defaultServiceManager();

LOGI("ServiceManager: %p", sm.get());

AudioFlinger::instantiate();

MediaPlayerService::instantiate();

CameraService::instantiate();

ProcessState::self()->startThreadPool();

IPCThreadState::self()->joinThreadPool();

}

AudioFlinger will call IServiceManager::addService, which is a RPC call.

void AudioFlinger::instantiate() {

defaultServiceManager()->addService(

String16("media.audio_flinger"), new AudioFlinger());

}

AudioFlinger inherits from BnAudioFlinger, which is a template BnInterface class.

class BnAudioFlinger : public BnInterface<IAudioFlinger>

{

public:

virtual status_t onTransact( uint32_t code,

const Parcel& data,

Parcel* reply,

uint32_t flags = 0);

};

BnInterface inherits from BBinder.

template<typename INTERFACE>

class BnInterface : public INTERFACE, public BBinder

{

public:

virtual sp<IInterface> queryLocalInterface(const String16& _descriptor);

virtual String16 getInterfaceDescriptor() const;

protected:

virtual IBinder* onAsBinder();

};

template<typename INTERFACE>

IBinder* BnInterface<INTERFACE>::onAsBinder()

{

return this;

}

According to BnInterface’m implementation we know that the parameter passed to

IServiceManager::addService is the address of new AudioFlinger instance. BBinder derives from

IBinder, its transact function is to call virtual function onTransact.

status_t BBinder::transact(

uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)

{

data.setDataPosition(0);

status_t err = NO_ERROR;

switch (code) {

case PING_TRANSACTION:

reply->writeInt32(pingBinder());

break;

default:

err = onTransact(code, data, reply, flags);

break;

}

if (reply != NULL) {

reply->setDataPosition(0);

}

return err;

}

The most important virtual functions is onTransact. BnAudioFlinger implemented the virtual

function. For the scenario, we just need to focus on SET_MODE branch.

status_t BnAudioFlinger::onTransact(

uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)

{

switch(code) {

case SET_MODE: {

CHECK_INTERFACE(IAudioFlinger, data, reply);

int mode = data.readInt32();

reply->writeInt32( setMode(mode) );

return NO_ERROR;

} break;

media_server will enter into a loop through IPCThreadState::joinThreadPool, just like

service_manager, it will be waited data from other process in talkWithDriver.

void IPCThreadState::joinThreadPool(bool isMain)

{

mOut.writeInt32(isMain ? BC_ENTER_LOOPER : BC_REGISTER_LOOPER);

status_t result;

do {

int32_t cmd;

result = talkWithDriver();

if (result >= NO_ERROR) {

size_t IN = mIn.dataAvail();

if (IN < sizeof(int32_t)) continue;

cmd = mIn.readInt32();

result = executeCommand(cmd);

}

// Let this thread exit the thread pool if it is no longer

// needed and it is not the main process thread.

if(result == TIMED_OUT && !isMain) {

break;

}

} while (result != -ECONNREFUSED && result != -EBADF);

mOut.writeInt32(BC_EXIT_LOOPER);

talkWithDriver(false);

}

Assume you want to implement your own service IFunnyTest, you must do the following:

 Implement your own BnFunnyTest class.

 In the process your service running, call IPCThreadState::joinThreadPool to start the loop for

binder.

RPC Call IServiceManager::addService

We called IServiceManager::addService, which means call BpServiceManager::addService.

virtual status_t addService(const String16& name, const sp<IBinder>& service)

{

Parcel data, reply;

data.writeInterfaceToken(IServiceManager::getInterfaceDescriptor());

data.writeString16(name);

data.writeStrongBinder(service);

status_t err = remote()->transact(ADD_SERVICE_TRANSACTION, data, &reply);

return err == NO_ERROR ? reply.readInt32() : err;

}

Parcel is simple. We can think it as a continuous buffer. Pay attention here, service parameter

points to BBinder object(AudioFlinger derived from Bn)

status_t Parcel::writeStrongBinder(const sp<IBinder>& val)

{

return flatten_binder(ProcessState::self(), val, this);

}

flatten_binder will generate a Binder command. Because BBinder is a local binder object, so the

code branch to the lines marked red.

status_t flatten_binder(const sp<ProcessState>& proc,

const sp<IBinder>& binder, Parcel* out)

{

flat_binder_object obj;

obj.flags = 0x7f | FLAT_BINDER_FLAG_ACCEPTS_FDS;

if (binder != NULL) {

IBinder *local = binder->localBinder();

if (!local) {

BpBinder *proxy = binder->remoteBinder();

if (proxy == NULL) {

LOGE("null proxy");

}

const int32_t handle = proxy ? proxy->handle() : 0;

obj.type = BINDER_TYPE_HANDLE;

obj.handle = handle;

obj.cookie = NULL;

} else {

obj.type = BINDER_TYPE_BINDER;

obj.binder = local->getWeakRefs();

obj.cookie = local;

}

}

return finish_flatten_binder(binder, obj, out);

}

Pay attention to the red lines, local’s address is put into the packet(It will be used later). After the

packet for addService RPC call is made, BpServiceManager::addService will call BpBinder’s

transact.

status_t BpBinder::transact(

uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)

{

// Once a binder has died, it will never come back to life.

if (mAlive) {

status_t status = IPCThreadState::self()->transact(

mHandle, code, data, reply, flags);

if (status == DEAD_OBJECT) mAlive = 0;

return status;

}

return DEAD_OBJECT;

}

BpBinder calls IPCThreadState::transact to start a transaction to the binder object corresponding

to mHandle. For this scenario, mHandle is 0.

status_t IPCThreadState::transact(int32_t handle,

uint32_t code, const Parcel& data,

Parcel* reply, uint32_t flags)

{

status_t err = data.errorCheck();

flags |= TF_ACCEPT_FDS;

if (err == NO_ERROR) {

LOG_ONEWAY(">>>> SEND from pid %d uid %d %s", getpid(), getuid(),

(flags & TF_ONE_WAY) == 0 ? "READ REPLY" : "ONE WAY");

err = writeTransactionData(BC_TRANSACTION, flags, handle, code, data, NULL);

}

if (err != NO_ERROR) {

if (reply) reply->setError(err);

return (mLastError = err);

}

if ((flags & TF_ONE_WAY) == 0) {

if (reply) {

err = waitForResponse(reply);

} else {

Parcel fakeReply;

err = waitForResponse(&fakeReply);

}

} else {

err = waitForResponse(NULL, NULL);

}

return err;

}

IPCThreadState::transact will first call writeTransactionData to compose a transaction structure

for binder kernel driver. Pay attention to the following lines, it’s very important for binder kernel

driver to identify the transaction target.

status_t IPCThreadState::writeTransactionData(int32_t cmd, uint32_t binderFlags,

int32_t handle, uint32_t code, const Parcel& data, status_t* statusBuffer)

{

binder_transaction_data tr;

tr.target.handle = handle;

tr.code = code;

tr.flags = binderFlags;

const status_t err = data.errorCheck();

if (err == NO_ERROR) {

tr.data_size = data.ipcDataSize();

tr.data.ptr.buffer = data.ipcData();

tr.offsets_size = data.ipcObjectsCount()*sizeof(size_t);

tr.data.ptr.offsets = data.ipcObjects();

} else if (statusBuffer) {

tr.flags |= TF_STATUS_CODE;

*statusBuffer = err;

tr.data_size = sizeof(status_t);

tr.data.ptr.buffer = statusBuffer;

tr.offsets_size = 0;

tr.data.ptr.offsets = NULL;

} else {

return (mLastError = err);

}

mOut.writeInt32(cmd);

mOut.write(&tr, sizeof(tr));

return NO_ERROR;

}

Then waitForResponse will call talkWithDriver to invoke BINDER_WRITE_READ ioctl.

#if defined(HAVE_ANDROID_OS)

if (ioctl(mProcess->mDriverFD, BINDER_WRITE_READ, &bwr) >= 0)

err = NO_ERROR;

else

err = -errno;

#else

So until now the transaction data has been delivered to binder kernel driver.

Summary:

Proxy object will generate the needed packet for a RPC call and then invoke

BINDER_WRITE_READ to write the packet to binder kernel driver. The packet is a formatted

packet. For RPC call, it uses BC_TRANSACTION packet type.

Assume you want to implement your own service IFunnyTest, you must do the following:

 In the process your service running, call IServiceManager::addService to register the service

to servier_manger.

Transaction in Binder Kernel Driver

When any process opens the “/dev/binder” driver, a corresponding structure binder_proc will be

assigned at binder_open.

static int binder_open(struct inode *nodp, struct file *filp)

{

struct binder_proc *proc;

proc = kzalloc(sizeof(*proc), GFP_KERNEL);

if (proc == NULL)

return -ENOMEM;

get_task_struct(current);

proc->tsk = current;

INIT_LIST_HEAD(&proc->todo);

init_waitqueue_head(&proc->wait);

proc->default_priority = task_nice(current);

mutex_lock(&binder_lock);

binder_stats.obj_created[BINDER_STAT_PROC]++;

hlist_add_head(&proc->proc_node, &binder_procs);

proc->pid= current->group_leader->pid;

INIT_LIST_HEAD(&proc->delivered_death);

filp->private_data

mutex_unlock(&binder_lock);

if (binder_proc_dir_entry_proc) {

char strbuf[11];

snprintf(strbuf, sizeof(strbuf), "%u", proc->pid);

create_proc_read_entry(strbuf,

binder_read_proc_proc, proc);

}

return 0;

}

S_IRUGO,

binder_proc_dir_entry_proc,

So that when any ioctl reached, the driver can know the process info. The transaction data is

transferred through BINDER_WRITE_READ ioctl.

case BINDER_WRITE_READ: {

struct binder_write_read bwr;

if (size != sizeof(struct binder_write_read)) {

ret = -EINVAL;

goto err;

}

if (copy_from_user(&bwr, ubuf, sizeof(bwr))) {

ret = -EFAULT;

goto err;

}

if (bwr.write_size > 0) {

ret = binder_thread_write(proc, thread, (void __user *)bwr.write_buffer,

bwr.write_size, &bwr.write_consumed);

if (ret < 0) {

bwr.read_consumed = 0;

if (copy_to_user(ubuf, &bwr, sizeof(bwr)))

ret = -EFAULT;

goto err;

}

}

if (bwr.read_size > 0) {

ret = binder_thread_read(proc, thread, (void __user *)bwr.read_buffer,

bwr.read_size, &bwr.read_consumed, filp->f_flags & O_NONBLOCK);

if (!list_empty(&proc->todo))

wake_up_interruptible(&proc->wait);

if (ret < 0) {

if (copy_to_user(ubuf, &bwr, sizeof(bwr)))

ret = -EFAULT;

goto err;

}

}

if (copy_to_user(ubuf, &bwr, sizeof(bwr))) {

ret = -EFAULT;

goto err;

}

break;

}

The driver first handles write, then read. Let’s take a look at the binder_thread_write first. The

core in binder_thread_write is a loop to parse command from write buffer and execute the

corresponding command.

uint32_tcmd;

void __user *ptr = buffer + *consumed;

void __user *end = buffer + size;

while (ptr < end && thread->return_error == BR_OK) {

if (get_user(cmd, (uint32_t __user *)ptr))

return -EFAULT;

ptr +=sizeof(uint32_t);

if (_IOC_NR(cmd) < ARRAY_SIZE(binder_stats.bc)) {

binder_stats.bc[_IOC_NR(cmd)]++;

proc->stats.bc[_IOC_NR(cmd)]++;

thread->stats.bc[_IOC_NR(cmd)]++;

}

switch (cmd) {

case ***:

default:

printk(KERN_ERR "binder: %d:%d unknown command %d\n", proc->pid,

thread->pid, cmd);

return -EINVAL;

}

*consumed = ptr - buffer;

}

We just take a look at 2 commands related to the scenario. One is BC_INCREFS.

case BC_INCREFS:

case BC_ACQUIRE:

case BC_RELEASE:

case BC_DECREFS: {

uint32_ttarget;

struct binder_ref *ref;

const char *debug_string;

if (get_user(target, (uint32_t __user *)ptr))

return -EFAULT;

ptr +=sizeof(uint32_t);

if (target == 0 && binder_context_mgr_node &&

(cmd == BC_INCREFS || cmd == BC_ACQUIRE)) {

ref = binder_get_ref_for_node(proc,

binder_context_mgr_node);

} else

ref = binder_get_ref(proc, target);

if (ref == NULL) {

binder_user_error("binder: %d:%d refcou"

"nt change oninvalid ref %d\n",

proc->pid,thread->pid, target);

break;

}

switch (cmd) {

case BC_INCREFS:

debug_string = "IncRefs";

binder_inc_ref(ref, 0, NULL);

break;

}

break;

Remember that, we mentioned it before, in this scenario, the target will be 0.

binder_context_mgr_node is created to represent handle 0 when system_manager calls

BINDER_SET_CONTEXT_MGR ioctl. So here just increase a weak reference to

binder_context_mgr_node node.

binder_context_mgr_node =binder_new_node(proc, NULL);

The other is BC_TRANSACTION.

case BC_TRANSACTION:

case BC_REPLY: {

struct binder_transaction_data

if (copy_from_user(&tr, ptr,sizeof(tr)))

return -EFAULT;

ptr += sizeof(tr);

binder_transaction(proc, thread, &tr, cmd == BC_REPLY);

break;

}

binder_transaction will create a new binder node if the packet contains a

BINDER_TYPE_BINDER flattened object.

fp = (struct flat_binder_object *)(t->buffer->data + *offp);

switch (fp->type) {

case BINDER_TYPE_BINDER:

case BINDER_TYPE_WEAK_BINDER: {

struct binder_ref *ref;

structbinder_node *node = binder_get_node(proc, fp->binder);

if (node == NULL) {

node =binder_new_node(proc, fp->binder);

if (node == NULL) {

return_error = BR_FAILED_REPLY;

goto err_binder_new_node_failed;

}

node->cookie = fp->cookie;

}

ref = binder_get_ref_for_node(target_proc, node);

if (ref == NULL) {

return_error = BR_FAILED_REPLY;

goto err_binder_get_ref_for_node_failed;

}

if (fp->type == BINDER_TYPE_BINDER)

fp->type = BINDER_TYPE_HANDLE;

else

fp->type = BINDER_TYPE_WEAK_HANDLE;

fp->handle = ref->desc;

binder_inc_ref(ref, fp->type == BINDER_TYPE_HANDLE, &thread->todo);

if (binder_debug_mask& BINDER_DEBUG_TRANSACTION)

printk(KERN_INFO " node %d u%p -> ref %d desc %d\n",

node->debug_id, node->ptr, ref->debug_id, ref->desc);

} break;

binder_transaction will know that the target is handle 0, so it runs the following branch to get the

target_node, target_proc and target_thread.

} else {

target_node = binder_context_mgr_node;

if (target_node == NULL) {

return_error = BR_DEAD_REPLY;

goto err_no_context_mgr_node;

}

}

e->to_node = target_node->debug_id;

target_proc = target_node->proc;

if (!(tr->flags & TF_ONE_WAY) && thread->transaction_stack) {

struct binder_transaction *tmp;

tmp = thread->transaction_stack;

while (tmp) {

if (tmp->from && tmp->from->proc == target_proc)

target_thread = tmp->from;

tmp = tmp->from_parent;

}

}

Finally binder_transaction will put this request into list and wake up the waiting thread in

binder_thread_read.

t->work.type = BINDER_WORK_TRANSACTION;

list_add_tail(&t->work.entry,target_list);

tcomplete->type = BINDER_WORK_TRANSACTION_COMPLETE;

list_add_tail(&tcomplete->entry, &thread->todo);

if (target_wait)

wake_up_interruptible(target_wait);

Now let’s take a look at binder_thread_read. When service_manager runs, it will wait at here until

a request is delivered to it.

ret = wait_event_interruptible_exclusive(proc->wait, binder_has_proc_work(proc,

thread));

Because the previous write from media_server process has wakened it up, so it gets executed. The

following code is to copy data from write buffer of media_server to read buffer of

system_manager.

tr.data_size = t->buffer->data_size;

tr.offsets_size = t->buffer->offsets_size;

tr.data.ptr.buffer = (void *)((void *)t->buffer->data + proc->user_buffer_offset);

tr.data.ptr.offsets = tr.data.ptr.buffer + ALIGN(t->buffer->data_size, sizeof(void *));

if (put_user(cmd, (uint32_t __user *)ptr))

return -EFAULT;

ptr +=sizeof(uint32_t);

if (copy_to_user(ptr, &tr, sizeof(tr)))

return -EFAULT;

ptr += sizeof(tr);

Summary:

This section has shown the data flow from RPC call client side to RPC server side.

Service Manager Handle Add Service

Until now, the service_manager has gotten a packet of BR_TRANSACTION from media_server,

so it will call binder_parser to handle the packet.

case BR_TRANSACTION: {

if (func) {

unsigned rdata[256/4];

struct binder_io msg;

struct binder_io reply;

int res;

bio_init(&reply, rdata, sizeof(rdata), 4);

bio_init_from_txn(&msg, txn);

res = func(bs, txn, &msg, &reply);

binder_send_reply(bs, &reply, txn->data, res);

}

ptr += sizeof(*txn) / sizeof(uint32_t);

break;

}

binder_parser will call svcmgr_handler to parse BR_TRANSACTION packet, which is the reverse

process of BpServerManager. Here structure binder_txn actually is the same with structure

binder_transaction_data. In our scenario, the transaction code is SVC_MGR_ADD_SERVICE.

int svcmgr_handler(struct binder_state *bs,

struct binder_txn *txn,

struct binder_io *msg,

struct binder_io *reply)

{

struct svcinfo *si;

uint16_t *s;

unsigned len;

void *ptr;

if (txn->target != svcmgr_handle)

return -1;

s = bio_get_string16(msg, &len);

if ((len != (sizeof(svcmgr_id) / 2)) ||

memcmp(svcmgr_id, s, sizeof(svcmgr_id))) {

fprintf(stderr,"invalid id %s\n", str8(s));

return -1;

}

switch(txn->code) {

case SVC_MGR_ADD_SERVICE:

s = bio_get_string16(msg, &len);

ptr = bio_get_ref(msg);

if (do_add_service(bs, s, len, ptr, txn->sender_euid))

return -1;

break;

So that service_manager knows that a service named s will run and it gets the object information

through bio_get_ref.

void *bio_get_ref(struct binder_io *bio)

{

struct binder_object *obj;

obj = _bio_get_obj(bio);

if (!obj)

return 0;

if (obj->type == BINDER_TYPE_HANDLE)

return obj->pointer;

return 0;

}

bio_get_ref is the reverse work of flatten_binder. do_add_service will finally call BC_ACQUIRE

to get a strong reference to the object represented by ptr.

Summary:

This section shows how the service is added to service manager.

Assume you want to implement your own service IFunnyTest, you must do the following:

 Add your service name into allowed service list in service_manager.

Get IAudioFlinger

The only way to get a service interface is through IServiceManager::getService. For example, here

is the method for AudioSystem to get an IAudioFlinger.

// establish binder interface to AudioFlinger service

const sp<IAudioFlinger>& AudioSystem::get_audio_flinger()

{

Mutex::Autolock _l(gLock);

if (gAudioFlinger.get() == 0) {

sp<IServiceManager> sm = defaultServiceManager();

sp<IBinder> binder;

do {

binder = sm->getService(String16("media.audio_flinger"));

if (binder != 0)

break;

LOGW("AudioFlinger not published, waiting...");

usleep(500000); // 0.5 s

} while(true);

gAudioFlinger = interface_cast<IAudioFlinger>(binder);

}

LOGE_IF(gAudioFlinger==0, "no AudioFlinger!?");

return gAudioFlinger;

}

IServiceManager::getService is called to BpServiceManager::getService.

virtual sp<IBinder> getService(const String16& name) const

{

unsigned n;

for (n = 0; n < 5; n++){

sp<IBinder> svc = checkService(name);

if (svc != NULL) return svc;

LOGI("Waiting for sevice %s...\n", String8(name).string());

sleep(1);

}

return NULL;

}

virtual sp<IBinder> checkService( const String16& name) const

{

Parcel data, reply;

data.writeInterfaceToken(IServiceManager::getInterfaceDescriptor());

data.writeString16(name);

remote()->transact(CHECK_SERVICE_TRANSACTION, data, &reply);

return reply.readStrongBinder();

}

Just like analyzed before, the call will finally through Binder kernel driver to be handled in

service_manager process.

switch(txn->code) {

case SVC_MGR_GET_SERVICE:

case SVC_MGR_CHECK_SERVICE:

s = bio_get_string16(msg, &len);

ptr = do_find_service(bs, s, len);

if (!ptr)

break;

bio_put_ref(reply, ptr);

return 0;

Then service_manager will reply with the previous handle set by media_server (Which is the

AudioFlinger’s instance’s address.). Then BpServiceManager::checkService will be returned back

from remote()->transact call. Then just like analyzed for IServiceManager, it will new another

BpBinder instance corresponding to the handle returned back from service_manager.

interface_cast<IAudioFlinger>(binder) will finally return an BpAudioFlinger instance.

Summary:

Just like get IServieManager, but this time it needs to get a handle from service_manager. While

for IServiceManager it always use handle 0.

RPC Call IAudioFlinger::SetMode

If we call IAudioFlinger::SetMode in AAA process, actually call BpAudioFlinger::setMode.

virtual status_t setMode(int mode)

{

Parcel data, reply;

data.writeInterfaceToken(IAudioFlinger::getInterfaceDescriptor());

data.writeInt32(mode);

remote()->transact(SET_MODE, data, &reply);

return reply.readInt32();

}

Like analysis of IServiceManager::addService, this function will generate a packet and write it to

binder kernel driver, then wait to read reply. The only difference is this time the target handle

points to some address in media_server process.

Handle IAudioFlinger::SetMode

The Binder kernel driver will finally wake up the read thread of media_server process which runs

at IPCThreadState::joinThreadPool. Now let’s re-look the code.

void IPCThreadState::joinThreadPool(bool isMain)

{

mOut.writeInt32(isMain ? BC_ENTER_LOOPER : BC_REGISTER_LOOPER);

status_t result;

do {

int32_t cmd;

result = talkWithDriver();

if (result >= NO_ERROR) {

size_t IN = mIn.dataAvail();

if (IN < sizeof(int32_t)) continue;

cmd = mIn.readInt32();

result = executeCommand(cmd);

}

// Let this thread exit the thread pool if it is no longer

// needed and it is not the main process thread.

if(result == TIMED_OUT && !isMain) {

break;

}

} while (result != -ECONNREFUSED && result != -EBADF);

mOut.writeInt32(BC_EXIT_LOOPER);

talkWithDriver(false);

}

At this time, talkWithDriver will return with the packet generated by BpServiceManager::setMode,

then executeCommand will process the commands. In this scenario, the command is

BR_TRANSACTION.

case BR_TRANSACTION:

{

binder_transaction_data tr;

Parcel reply;

if (tr.target.ptr) {

sp<BBinder> b((BBinder*)tr.cookie);

const status_t error = b->transact(tr.code, buffer, &reply, 0);

if (error < NO_ERROR) reply.setError(error);

} else {

const status_t error = the_context_object->transact(tr.code, buffer, &reply, 0);

if (error < NO_ERROR) reply.setError(error);

}

if ((tr.flags & TF_ONE_WAY) == 0) {

LOG_ONEWAY("Sending reply to %d!", mCallingPid);

sendReply(reply, 0);

} else {

LOG_ONEWAY("NOT sending reply to %d!", mCallingPid);

}

}

break;

The most important two lines are marked with red. Here it gets an address from binder kernel

driver and cast it to a BBinder pointer(This address in put into binder kernel driver when called

IServiceManager::addService.). Remember that, AudioFlinger is derived from BBinder. This

pointer is actually the same pointer of our AudioFlinger instance. So the following transact call

will finally call our BnAudioFlinger’s onTransact virtual function.

status_t BnAudioFlinger::onTransact(

uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)

{

case SET_MODE: {

CHECK_INTERFACE(IAudioFlinger, data, reply);

int mode = data.readInt32();

reply->writeInt32( setMode(mode) );

return NO_ERROR;

} break;

}

Then the reply will be written through sendReply.

status_t IPCThreadState::sendReply(const Parcel& reply, uint32_t flags)

{

status_t err;

status_t statusBuffer;

err = writeTransactionData(BC_REPLY, flags, -1, 0, reply, &statusBuffer);

if (err < NO_ERROR) return err;

return waitForResponse(NULL, NULL);

}

It will finally write to binder kernel driver. Then kernel driver will wake up the read thread of

AAA.

Summary

The overall architecture of Android IPC system is shown in the picture. There are four major

blocks:

 Binder Driver

It’s the core of IPC system. It passes data between service provider and service user.

 service provider

It provides some kind of service. It will parser the received RPC call data from binder driver

and do the real action.

 service_manager

It’s a special service provider. It provides service manager service for other service provider.

 service user

It remote calls service provider. It will generate an RPC call data and send it to binder driver.

For the example scenario, here lists the major control flow.

1. service_manager runs first, it will register a special node 0 in binder driver.

2. media_server gets an IServiceManager proxy object for the special node 0.

3. media_server RPC call IServiceManager::addService to add the IAudioFlinger service. This

call is routed to node 0. It will send data to binder driver.

4. binder driver notify that data is for node 0, and the data contains command to binder a object.

So it will generate another node (assume A) for the IAudioFlinger service and route the data

to service_manager.

5. service_manager reads data from binder driver, then process IServiceManager::addService

RPC call.

6. another process P gets an IServiceManager proxy object for the special node 0.

7. P RPC call IServiceManager::getService to get IAudioFlinger service. This call is routed to

node 0. It will send data to binder driver.

8. binder driver notify that data is for node 0, So it will route the data to service_manager.

9. service_manager reads data from binder driver, then process IServiceManager::getService

RPC call and return back the node A to represent the IAudioFlinger service.

10. P RPC call IAudioFlinger::setMode. Now this call will be routed to node A.

11. binder driver notify that data is for node A, so it will route the data to media_server.

12. media_server reads data from binder driver, handles IAudioFlinger::setMode RPC call and

send reply data to binder driver.

13. binder driver route the reply to P.

14. P reads data from binder driver and finally gets the reply.

更多相关文章

  1. 代码中设置drawableleft
  2. android 3.0 隐藏 系统标题栏
  3. Android开发中activity切换动画的实现
  4. Android(安卓)学习 笔记_05. 文件下载
  5. Android中直播视频技术探究之—摄像头Camera视频源数据采集解析
  6. 技术博客汇总
  7. android 2.3 wifi (一)
  8. AndRoid Notification的清空和修改
  9. Android中的Chronometer

随机推荐

  1. Android第二十七课 NDK 渲染色彩深度
  2. Android(安卓)支持多屏幕机制
  3. 巧妙利用PARTITION分组排名递增特性解决
  4. sql编程工具Sql Prompt下载及安装破解图
  5. 如何使用Visual Studio 2010在数据库中生
  6. MSSQL事务的存储过程
  7. SQL Server UPDATE语句的用法详解
  8. sql存储过程详解
  9. 总结一周内学习的Sql经验(一)
  10. SQL SERVER使用REPLACE将某一列字段中的